Inactivation of p53 tumor suppressor gene acts synergistically ... - Nature

2 downloads 57 Views 617KB Size Report
Inactivation of p53 tumor suppressor gene acts synergistically with c-neu oncogene in salivary gland tumorigenesis. Steven G Brodie1, Xiaoling Xu1, Cuiling Li1, ...
Oncogene (2001) 20, 1445 ± 1454 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Inactivation of p53 tumor suppressor gene acts synergistically with c-neu oncogene in salivary gland tumorigenesis Steven G Brodie1, Xiaoling Xu1, Cuiling Li1, Ann Kuo2, Philip Leder2 and Chu-Xia Deng*,1 1

Genetics of Development and Disease Branch, 10/9N105, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland, MD 20892, USA; 2Department of Genetics, Harvard Medical School, Howard Hughes Medical Institute, 200 Longwood Avenue, Boston, Massachusetts, MA 02115, USA

Transgenic mice expressing speci®c oncogenes usually develop tumors in a stochastic fashion suggesting that tumor progression is a multi-step process. To gain further understanding of the interactions between oncogenes and tumor suppressor genes during tumorigenesis, we have crossed a transgenic strain (TG.NK) carrying an activated c-neu oncogene driven by the MMTV enhancer/promoter with p53-de®cient mice. cneu transgenic mice have stochastic breast tumor formation and normal appearing salivary glands. However, c-neu mice heterozygous for a p53 deletion develop parotid gland tumors and loose their wild type p53 allele. c-neu mice with a homozygous p53 deletion have increased rates of parotid tumor onset suggesting that inactivation of p53 is required and sucient for parotid gland transformation in the presence of activated c-neu. In contrast to the dramatic e€ect of p53 in parotid gland transformation, p53 loss has little e€ect on the rate or stochastic appearance of mammary tumors. In addition, p53 loss was accompanied by the down regulation of p21 in parotid gland tumors but not breast tumors. The parotid gland tumors were aneuploid and demonstrated increased levels of Cyclin D1 expression. These observations suggest that in c-neu transgenic mice, p53 alterations have di€erential tissue e€ects and may be in¯uenced by the tissue speci®c expression of genes in¯uencing p53 activity. Oncogene (2001) 20, 1445 ± 1454. Keywords: p53; erbB2; Cyclin cooperation; parotid gland

D1;

tumorigenesis;

Introduction The most frequent genetic alteration found in breast cancer is the ampli®cation or overexpression of c-erbB2 (c-neu, HER2) (reviewed in Menard et al., 2000). Elevated expression of c-erbB2 has been identi®ed in 30 ± 40% of human breast cancers and is associated with a poor prognosis (Slamon et al., 1987). Subsequent to these ®ndings, c-erbB2 ampli®cation has been

*Correspondence: C-X Deng Received 11 October 2000; revised 21 December 2000; accepted 4 January 2001

identi®ed in a variety of tumors including ovarian, pancreatic and salivary gland tumors (Hall et al., 1990; Muller et al., 1994; Slamon et al., 1989). The c-erbB2 proto-oncogene is one member of the epidermal growth factor receptor (EGFR) family that includes EGFR (erbB1), erbB3 and erbB4 (Earp et al., 1995; Hynes, 1996; Menard et al., 2000). These receptor tyrosine kinases bind a number of EGF-ligands leading to hetrodimerization, autophosphorylation and receptor activation (Graus-Porta et al., 1997; Pinkas-Kramarski et al., 1996, 1997; Tzahar et al., 1996). An activated form of the c-neu proto-oncogene was originally isolated from chemically induced rat neuroblastoma cells. This activating mutation was (Shih et al., 1981) caused by a single aminoacid substitution (Val-664-Glu) in its transmembrane domain (Bargmann et al., 1986). Notably, transgenic mice expressing this activating mutation develop rapid multi-focal breast tumors (TG.NF strain) or formed slower growing stochastic breast tumors (TG.NK strain) (Muller et al., 1988). Analysis of these mice suggests that activated c-neu could eciently transform mammary gland epithelium in a single step and that its in¯uence was dependent upon a `positional e€ect' of the integration site. One interesting ®nding was that in transgenic (TG.NF strain) mice where c-neu is driven by the MMTV promotor the gene is also expressed in the epididymis, salivary and harderian glands. However, only benign hyperplasia was observed in these tissues suggesting tissue speci®c modi®ers may confer resistance to transformation by activated c-neu (Bouchard et al., 1989; Lucchini et al., 1992; Muller et al., 1988). The overall results of these studies suggest that in addition to c-neu activation, tissues other than the mammary epithelium require tissue speci®c modi®ers or secondary genetic changes for tumor formation. In humans, approximately half of all tumors carry p53 mutations, suggesting that p53 plays a central role in tumor progression (Morgan and Kastan, 1997; Vogelstein and Kinzler, 1992). This tumor suppressor acts with the induction of p21, as a cell cycle checkpoint, arresting cells in G1/S in response to DNA damage, preventing proliferation (Deng et al., 1995; Lee and Bernstein, 1993; Linke et al., 1997; reviewed in Levine, 1997). p53 represses tumor progression through the induction of programmed cell death (apoptosis). This processes involves activation of

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1446

target genes such as bcl-2 and bax leading to a G2/M cell cycle block and subsequent cell death (el-Deiry et al., 1993; Hermeking et al., 1997; Knudson and Korsmeyer, 1997; Knudson et al., 1995; Polyak et al., 1997; Yin et al., 1999). These roles of p53 are critical for maintaining cellular integrity and preventing tumor formation. In addition to p53 mutations causing somatic cancers, germline mutations have also been identi®ed in Li-Fraumeni syndrome, a genetic disease predisposing individuals to cancer (Malkin et al., 1992; Tonin, 2000; Varley et al., 1997). Also, in vivo studies using p53-de®cient mice have recapitulated this genetic disease and provides further evidence for its role in the formation of a variety of tumors (Donehower et al., 1992; Jacks et al., 1994). Taking advantage of the presence of p53-de®cient mice generated by gene targeting (Donehower et al., 1992), we investigated the cooperation between p53 and the c-neu proto-oncogene in tumor formation. By generating mice carrying an activated form of c-neu and a targeted deletion of p53, we were interested in testing whether the genes collaborate to cause transformation of tissues where the transgene is expressed. In c-neu transgenic mice, p53 loss triggered the formation of parotid gland tumors but had little e€ect on the rate or stochastic appearance of breast tumors. In addition, loss of p53 in parotid gland tumors also caused down regulation of one of its downstream genes, p21. However, in breast tumors, p53 loss had no e€ect on p21 transcription. The parotid gland tumors were aneuploid and had increased Cyclin D1 expression. Our results demonstrate that p53 inactivation in c-neu transgenic mice has di€erential e€ects on tumorigenesis in di€erent tissues.

Results Salivary tumors form in c-neu transgenic mice in cooperation with p53 Previous studies have demonstrated that a variety of spontaneous tumors develop rapidly in p53-de®cient mice (Donehower et al., 1992). Therefore, we initially introduced a single deleted p53 allele into c-neu transgenic mice (TG.NK strain) in order to generate three genotypes (c-neu;p53+/7, p53+/7 and c-neu;p53+/+). The c-neu;p53+/7 mice appeared phenotypically normal with respect to growth rate, fertility and tissue morphology. However, by 16 weeks of age, large masses began to form in their salivary glands. With increasing age, unilateral and/or bilateral salivary tumors were observed in the majority of the cneu;p53+/7 mice (34/38) by 35 weeks (Figure 1). The salivary glands of the four remaining mice appeared morphologically normal. However, histologic examination showed multiple foci of transformed cells in their salivary gland. In contrast to the salivary tumors observed in c-neu;p53+/7 mice, the c-neu;p53+/+ (n=34) or p537/7 animals (n=15) never developed salivary tumors when analysed up to 18 months of age Oncogene

Figure 1 Salivary gland tumor formation and distribution found in MMTV-c-neu mice carring various p53 genotypes. Survival curves showing the percentage of salivary tumor-free mice as a function of time for c-neu;p53+/7, c-neu;p53+/+, c-neu;p537/7 and p537/7 mice. (n=number of mice; T=time)

(Figure 1). Some of the p53+/7 mice developed other types of spontaneous tumors that were similar to tumors described previously (Donehower et al., 1992; Harvey et al., 1993). Because p53 heterozygous mice or c-neu animals carrying wild type p53 alleles never developed salivary gland tumors, it is likely the tumors found in c-neu;p53+/7 mice result from the additive e€ects of the activated c-neu oncogene and the p53 deletion. Inactivation of p53 leads to rapid salivary gland tumor formation in c-neu transgenic mice Next, we determined the status of the wild type p53 allele in the salivary tumors found in c-neu;p53+/7 mice. As shown in Figure 2, Southern blot analysis demonstrated that 19 of 20 salivary tumors had lost their wild type p53 allele (Figure 2a,b). This data would suggest that elimination of the wild type p53 allele is required for salivary gland tumor formation in c-neu transgenic mice. However, an alternative explanation could be that loss of p53 is a secondary event accompanied by the growth of transformed cells. To distinguish these possibilities, the rate of salivary tumor formation was accessed in c-neu;p537/7 mice (Figure 1). If loss of the wild type p53 allele is required for salivary gland tumor formation, we hypothesized that introducing the c-neu oncogene into a p53-de®cient background would increase the rate of salivary gland tumor formation. On the other hand, if p53 loss were a secondary consequence, then the rate of salivary tumor formation would remain unchanged between c-neu transgenic mice carrying a single copy (c-neu;p53+/7) or no copies (c-neu;p537/7) of p53. As shown in Figure 1, salivary gland tumor formation in c-neu mice lacking p53 (c-neu;p537/7) was greatly accelerated and all of the mice developed bilateral or unilateral salivary tumors by 16 weeks, many of which were visibly detectable as

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

(p53+/+, p53+/7, p537/7) using a probe speci®c to MMTV-c-neu transgene (Figure 2c). As shown in Figure 2d, abundant c-neu transcripts were detectable in all salivary gland tumors examined (n=15). This result suggests that salivary gland tumorigenesis requires the synergetic activation of c-neu and p53 inactivation. Surprisingly, the expression of the c-neu transgene is detectable in normal glands only after long exposures suggesting low expression levels of the oncogene in normal salivary glands (data not shown).

1447

Salivary tumors are derived from foci of transformed cells in the parotid gland

Figure 2 Loss of heterozygosity (LOH) at the p53 locus in salivary gland tumors. (a) p53 construct. The p53 gene is ¯anked by two EcoRI sites approximately 16 kb apart. A neo gene insertion containing an EcoRI site between the endogenous sites generates two novel restriction fragments of about 8.5 kb. Hybridization with the full length p53 cDNA probe detects a 16 kb fragment in wild type mice (c-neu;p53+/+), the 16 kb fragment plus a shifted 8.5 kb fragment in heterozygous mice (cneu;p53+/7), and two 8.5 kb fragments in the homozygous mice (c-neu;p537/7). (b) LOH analysis from seven salivary tumors from c-neu;p53+/7 mice (lanes 1, 3, 4, 5, 6, 7 and 9). Lanes 2 and 8, spleen samples from a c-neu;p53+/7 mouse (arrowheads point to pseudogene fragments). Southern blot analysis of adjacent normal salivary tissue revealed no LOH (not shown). (c) MMTVc-neu oncogene construct. (d) Northern blot using a 32P-labeled SV40 poly (A)+ fragment probe and reprobed with a 300 bp 32Plabeled L32 fragment as loading control. RNA isolated from salivary tumors (lanes 1 ± 4), and breast tumors (lanes 5 and 6) from c-neu;p53+/7 mice. RNA isolated from wild type salivary glands (lanes 7 and 8)

early as 8 weeks of age. These results suggest that loss of the wild type p53 allele is an important factor in salivary gland tumor formation in c-neu transgenic mice. Therefore, p53 inactivation is required for salivary gland tumor formation in c-neu transgenic mice. c-neu expression in salivary glands and tumors Our initial experiments established a crucial role for p53 in salivary gland tumor formation in c-neu transgenic mice. Since mice carrying only the c-neu transgene do not develop salivary gland tumors, we asked whether the salivary tumors expressed the c-neu transgene. Northern analysis was performed on representative samples of normal salivary glands and/or tumors from c-neu transgenic mice carrying the three p53 genotypes

The salivary gland is divided into the major glands and numerous minor accessory glands that are found throughout the oral mucosa. The major salivary gland is further divided into the parotid, sublingual and submandibular glands (Figure 3a ± d). Salivary gland tumors account for less than 2% of all cancers and they have been identi®ed generating from both the major and minor salivary glands (Press et al., 1994). Therefore, we were interested in determining speci®cally which tissues in the salivary gland generated the tumors in our c-neu;p53+/7 mice. The salivary glands from cneu;p53+/7, c-neu;p53+/+, and c-neu;p537/7 mice were dissected, ®xed and sectioned for histologic examination. The salivary glands of c-neu;p53+/+ mice had normal morphology with no obvious histologic changes in the glands (n420). In contrast, histologic analysis of salivary glands of ®ve c-neu;p53+/7 mice at stages prior to salivary tumor formation revealed multiple loci of transformed cells in their parotid glands (Figure 3e). With time, these foci enlarged and quickly developed into solid tumor (Figure 3f,g). Next, we performed in situ hybridization studies using a c-neu antisence probe and found no detectable levels of message in c-neu;p53+/+ mice. However, one 12-month old c-neu mouse had foci of cells in the parotid gland with enlarged nuclei and strong expression of c-neu. Although the status of the foci cannot be determined, we predict the foci would not lead to a salivary tumor since no salivary tumors were identi®ed in c-neu;p53+/+ mice analysed between 6 ± 12 months (n=30) or 1 ± 1.5 years (n=15) of age. In contrast to the normal appearing salivary glands of c-neu transgenic mice, the c-neu;p537/7 (Figure 4a,b) and c-neu;p53+/7 animals (Figure 4c,d) developed multiple foci in their parotid glands with strong c-neu expression. These transformed foci could be observed when the transgenic mice were two (c-neu;p537/7) and ®ve (c-neu;p53+/7) months of age. These results would suggest that the salivary tumors found in c-neu;p53+/7 and c-neu;p537/7 mice generate from multiple foci throughout the parotid gland. p53 inactivation does not increase the rate of breast tumor formation in c-neu transgenic mice As originally reported, c-neu oncogene expression in the mammary glands of female mice is non-uniform Oncogene

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1448

our analysis focused on breast tumors in c-neu;p53+/7 female mice. Comparing the tumor kinetics of cneu;p53+/+ and c-neu;p53+/7 mice, the mean age of breast tumor formation in the latter was slightly shortened from 30 to 25 weeks (Figure 5a). The appearance of breast tumors in both groups appeared stochastic and monoclonal in nature. Southern analysis of c-neu;p53+/7 breast tumors indicated that 20% had lost their wild type p53 allele (n=30) (Figure 5b). This loss of heterozygosity is consistent with the slightly enhanced tumor formation (Figure 5a). In contrast to the dramatic tumor formation observed in salivary glands upon p53 deletion, our data suggests that loss of p53 has minimal e€ect upon breast tumor formation in c-neu transgenic mice. Expression of p53 and p21 in salivary and mammary gland tumors

Figure 3 Histologic characteristics of the major salivary glands. (a) H&E stained normal salivary gland from an 8 month old wild type mouse (parotid gland (P) consists primarily of serous cells, sublingual gland (Sl) consists primarily of mucous cells and submandibular gland (Sm) consists of both cell types). Histologic characteristics of the parotid (b), sublingual (c) and submandibular (d) glands of wild type mice. Parotid gland with tumor foci formation in a 5 month old c-neu;p53+/7 mouse (e) Arrows point to regions of foci. (f,g) Salivary tumor from an 8 month old cneu;p53+/7 mouse. Note the large foci (f) (arrow) surrounded by a sheet of tumor cells (f,g). All samples are H&E stained. Magni®cations: 106 for b ± d; 206 for e ± g

and mammary carcinomas develop in a sporadic and stochastic fashion (Muller et al., 1988). In our study, the majority of c-neu female mice developed tumors only after multiple pregnancies and our c-neu male mice never developed breast tumors. The breast tumors were ®rst detected in c-neu;p53+/+ mice at 6 months and by 12 months, over 86% (12 of 14) of female mice developed mammary carcinomas (Figure 5a). The numbers of mammary tumors varied between mice (1 ± 8) with an average of four tumors per mouse. Next, we wanted to determine the e€ects of p53 deletions on the rate of breast tumor formation in c-neu;p537/7 and c-neu;p53+/7 mice. However, all of our c-neu;p537/7 females died by 4 months of age primarily caused by salivary tumors in addition to many other spontaneous tumors which appear to be caused by the p53 deletion itself (Table 1 and Donehower et al., 1992). Therefore, Oncogene

Cellular integrity is maintained by the various roles of p53 in cell cycle checkpoint control, apoptosis and genome stability. Perturbations in any of these functions may lead to tumorigenesis. Control of the cell cycle during the G1 phase by p53 is primarily through p21 (a potent inhibitor of cyclin dependent kinases (CDKs)), as previously demonstrated using p21-de®cient MEF cells that loose this checkpoint (Brugarolas et al., 1995; Deng et al., 1995). Therefore, we were interested in determining the status of p53 and p21 in the normal salivary and mammary glands and breast tumors of c-neu transgenic mice, in addition to the salivary tumors found in c-neu;p53+/7 mice. As shown in Figure 6a, Northern analysis demonstrates that p53 transcripts were detectable in all tissues except for salivary tumors. The presence of p53 transcripts in breast tumors (n=4) suggests that p53 is not involved in the repression of breast tumor progression in c-neu transgenic mice. Using p21 as a probe, detectable levels of transcript were found in normal salivary glands (n=4). In salivary tumors (n=19), p21 transcripts levels were signi®cantly decreased (Figure 6a and not shown). This data suggests that down regulation of p21 maybe correlated with the triggered loss of p53 in salivary gland tumors. This is in contrast to breast tumors where we detected an approximate threefold increase in p21 expression in four of seven breast tumors. Surprisingly, increased p21 expression was also detected in other tumors isolated from c-neu;p537/7 (Figure 6a). Because the salivary tumors from our c-neu;p53+/7 mice showed low levels of p21 expression, we were interested in determining whether down regulation of p21 itself could be responsible for the formation of parotid gland tumors. Therefore, we mated c-neu transgenic mice with mice de®cient in p21 (Deng et al., 1995) to determine its role in parotid gland tumor formation. As previously reported, p21 de®cient mice develop normally but have a loss of G1 checkpoint control (Deng et al., 1995). These mice (c-neu;p217/7) developed mammary tumors similar to c-neu transgenic mice in age of onset and frequency (not shown).

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1449

Figure 4 In situ hybridization analysis of the parotid gland. Light ®eld: (a), dark ®eld: (b) of enhanced focus staining in the parotid gland from a 2 month old c-neu;p537/7 animal as revealed by a 32P-labeled SV40 poly(A)+ probe. (c) Dark ®eld view of multiple foci in the parotid gland of a 5 month old c-neu;p53+/7 animal. (d) Dark ®eld view of a salivary gland tumor from a 5 month cneu;p53+/7 animal

However, tumors of the parotid gland were never identi®ed in these mice (n=22). Because the cneu;p217/7 mice did not develop parotid gland tumors, this suggests that p53 loss may be the initiating factor in salivary gland tumorigenesis and that salivary tumor formation does not act directly through p21. Aneuploidy and extensive tumor heterogeneity in p53-difficient tumors The cell cycle is tightly regulated by a number of proteins responsible for preventing the progression of damaged cells through the G1/S checkpoint (Dasika et al., 1999). This loss of cell cycle control usually occurs indirectly with the loss or mutations in p53. Genetic instability has been observed in cells and tumors from p53 de®cient mice (Donehower, 1997; Donehower et al., 1992). It has been suggested that the loss of p53 function may cause an overall loss of genome integrity that manifests as enhanced karyotypic and molecular changes. In addition, tumors generated by the ampli®cation of c-neu have demonstrated multiple genetic alterations including gene ampli®cation and aneuploidy (Isola et al., 1999; Jimenez et al., 2000; Stal et al., 1994). This would suggest that activation of cneu can also lead to multiple chromosome anomalies. Therefore, we were interested in determining the chromosome status of mammary (n=5) and salivary gland (n=5) tumors generated from c-neu;p53+/7 mice. Metaphase spreads from both the parotid gland and

breast tumors were aneuploid and nearly all metaphases analysed contained di€erent chromosome numbers suggesting extensive tumor heterogeneity (Figure 6b). Cyclin D1 is overexpressed in parotid gland tumors Tumor progression may involve increased proliferation and/or decreased apoptosis. One important component involved in the G1/S cell cycle checkpoint control is Cyclin D1 and its overexpression is found in a variety of tumors. Overexpression of Cyclin D1 is thought to inactivate Rb through hyperphosphorylation and trigger E2F release resulting in increased cellular proliferation (Oswald et al., 1994; Resnitzky and Reed, 1995; Schulze et al., 1994; Sherr, 1996). Recently, researchers demonstrated that Cyclin D1 is required by activated c-neu to produce a transformed phenotype (Lee et al., 2000). Their results suggested that c-neu activation of E2F mediates increased Cyclin D1 expression in breast tumors of c-neu transgenic mice. Therefore, we were interested in determining whether activated c-neu would cause increased Cyclin D1 expression in the parotid gland tumors and breast tumors from our c-neu;p53+/7 mice. Using immunohistochemistry, we found that Cyclin D1 was overexpressed in the mammary tumors (Figure 6c) and parotid gland tumors from c-neu;p53+/7 (n=5) mice (Figure 6d,e). This data suggests that increased proliferation caused by increased Cyclin D1 expression Oncogene

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1450

Figure 5 Breast tumor formation and LOH analysis of mammary tumors. (a) Survival curves showing percentage of mammary tumor free mice as a function of time for c-neu;p53+/7 and c-neu;p53+/+ female mice. (n=number of mice; T=time). (b) LOH analysis of 12 mammary tumors from c-neu;p53+/7 mice (lanes 2 ± 13). Spleen samples from a c-neu;p53+/7 mouse (lane 1) (arrowheads point to pseudogene fragments)

is one factor involved in both the mammary and parotid tumor progression. We also performed TUNEL assays to check for apoptosis in the salivary glands of 4 month old wild type, c-neu;p53+/+ and cneu;p53+/7 mice. However, no detectable di€erences were observed between the mice. Therefore, Cyclin D1 overexpression in conjunction with the loss of p53 would suggest that progression through the G1 phase of the cell cycle is increased in the parotid gland which leads to overproliferation of epithelial cells leading to tumor formation. Discussion Transgenic mice bearing oncogenes have been invaluable tools for in vivo studies into the mechanisms of tumorigenesis (Cardi€ et al., 1991; Webster and Muller, 1994). Most tumors generated in these mice arise in a stochastic fashion and appear monoclonal or oligoclonal in origin, suggesting they are derived from single or small numbers of cells in targeted tissues. Results of combinationial studies have suggested that oncogene expression alone is not sucient for malignant transformation in the majority of cells where the Oncogene

proto-oncogenes are expressed (Cardi€ et al., 1991; Hunter, 1991). In an e€ort to identify complementary factors involved in tumor progression, we have crossed the MMTV-c-neu transgenic mice with p53-de®cient mice. Our data showed that expression of c-neu or p53 loss alone in the salivary gland could not cause salivary tumors. However, the combination of c-neu expression and p53 loss dramatically promoted parotid gland tumor formation. In addition, tumor progression was also accompanied by cyclin D1 overexpression and aneuploidy. The synergistic interaction between c-neu and p53 was speci®c to the parotid gland and did not signi®cantly accelerate mammary gland formation or yield tumors in other tissues. Similar ®ndings were made between c-myc and p53 where p53 accelerated formation of T-cell lymphomas but not mammary tumorigenesis in MMTV-c-myc;p53+/7 mice (Elson et al., 1995). These studies provide excellent examples that tissue speci®c modi®cations can ultimately a€ect tumorigenesis. The salivary gland consists of the major glands, including the parotid, submandibular and sublingual and numerous smaller glands that are termed minor (Califano and Eisele, 1999; Hand et al., 1999). In humans, the majority of salivary gland tumors are generated in the parotid gland and most are benign. Overall, salivary tumors are infrequent and account for less than 2% of all tumors. However, the identi®cation of genetic factors involved in salivary gland tumorigenesis remains important for the management and treatment of these solid tumors (Rice, 1999). To this end, researchers have used a variety of transgenic approaches to generate models for studying salivary gland tumorigenesis (Samuelson, 1996). The transforming ability of the c-neu oncogene is greatly in¯uenced by several factors including the positional e€ects caused by the genome integration site, the background di€erences caused by tissue speci®c modi®er genes and/or the involvement of other complementary/inhibitor factors such as oncogenes or tumor suppressor genes (Bouchard et al., 1989; Lucchini et al., 1992; Muller et al., 1988). To identify speci®c tissues in¯uenced by the interactions between the c-neu oncogene and the p53 tumor suppressor during tumorigenesis, we crossed a c-neu transgenic strain (TG.NK) with p53-de®cient mice. We found that c-neu transgenic mice never developed parotid gland tumors. However, c-neu mice lacking one p53 allele developed parotid gland tumors that showed LOH for the remaining wild type allele. The cooperation between p53 and c-neu was further demonstrated by the accelerated parotid tumor formation in mice lacking both copies of p53 and the demonstration the c-neu was indeed expressed in the salivary gland. In humans, immunohistochemical studies have demonstrated that mutant p53 and erbb2 are overexpressed in salivary gland tumors (Kamio, 1996). Therefore, this data provides direct in vivo evidence for the synergistic e€ects of p53 and c-neu in salivary gland tumor formation. In contrast to parotid gland tumor progression, p53 alterations in c-neu transgenic mice did not signi®cantly

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1451

Table 1 Tumors identi®ed in p53 de®cient mice Number

Sex

Genotype

Age (weeks)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

M M M M M M M M M F F F M M M M F M M F M M F M M F M M M

7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 7/7 +/7 +/7

4 4 5 10 10 10 11 14 16 12 12 14 11 14 14 11 9 13 10 14 9 16 12 7 21 18 16 30 52

change the rate of mammary tumor formation in female mice. This result is consistent with our observation that only 20% of breast tumors from cneu;p53+/7 mice demonstrated loss of the wild type p53 allele. However, in parotid gland tumors, 95% of the tumors had LOH for p53. This would suggest that p53 loss in di€erent tissues has varied e€ects on tumor formation. One explanation for the lack of p53 repression in breast tumors of c-neu transgenic mice is that expression is absent in tissues transformed by the MMTV-c-neu oncogene. However, p53 transcripts were detected in the mammary gland tumors of c-neu mice excluding the possibility that p53 is not expressed in cells transformed by MMTV-c-neu. In addition, transcripts for p21, a downsteam gene of p53, was also detected in the normal mammary and breast tumors from c-neu transgenic mice. On the other hand, p21 transcripts were very low in the parotid gland tumors. These results suggest that p21 expression is tightly regulated by p53 in the parotid gland and p53 deletions leads to decreased levels of p21 expression in parotid tumors. One additional ®nding was that the majority of breast tumors from c-neu transgenic mice have increased levels of p21 expression, which was also observed in other types of tumors found in p53de®cient mice (Figure 6a). These observations would suggest that p53 and p21 regulation in c-neu transgenic mice may vary between mammary and parotid gland tumors. However, to state the role that p21 plays in parotid gland tumor formation based on our preliminary observations would be premature. However,

Tumor type

Anatomic site

Teratoma Teratoma Teratoma Spindle cell sarcoma Spindle cell sarcoma Malignant teratoma with choriocarcinoma Nodular hermangioma Spindle cell sarcoma Malignant lymphoma, hemangiosarcoma Spindle cell tumor Nodular hemangiosarcoma Spindle cell sarcoma Hemangiosarcoma Malignant lymphoma, hemangiosarcoma Malignant lymphoma Poorly differentiated sarcoma Spindle cell tumor Teratoma Necrotic mass Papillarysalivary gland tumor Lymphoma, spindle cell tumor Hemangiosarcoma Spindle cell sarcoma Spindle cell sarcoma Malignant lymphoma Malignant lymphoma Malignant lymphoma Malignant lymphoma Immunoblastic lymphoma

Testes Testes Testes Subcutis Subcutis Testis Skin Leg Spleen, leg Skin Subcutis Lymph node, neck Leg Thymus, skin Thymus Femur Skeletal muscle Testis Testis Salivary gland Thymus, ear Skin Mammary gland Skin Thymus Generalized Generalized Generalized Fat

c-neu;p217/7 mice did not demonstrate parotid gland tumors, which suggests that p21 de®ciency may not activate salivary gland tumorigenesis. Overall, our data would suggest that the tissue speci®c di€erences may be caused by the partial or complete redundancy of tumor suppressors in mammary gland. In addition to a role for p53 in cell cycle checkpoint control through the induction of p21, it also has roles in apoptosis and genome stability (Levine, 1997). Therefore, we tested the parotid glands for apoptosis (Yin et al., 1999). However, using TUNEL assays, no detectable di€erences were observed in the parotid glands of c-neu;p53+/+ and c-neu;p53+/7 mice at 5 months of age. This may be explained by the fact that the foci formed in the parotid glands are sparse making the detection of apoptosis in the foci dicult. Because tumorigenesis can also be caused by increased proliferation and Cyclin D1 has been demonstrated as being down stream of c-neu (Lee et al., 2000), we investigated whether Cyclin D1 was overexpressed in the parotid tumors of c-neu;p53+/7 mice. We found that the parotid gland tumors of c-neu;p53+/7 mice had increased levels of Cyclin D1 expression. These results are consistent with recent results in breast tumors from MMTV-c-neu transgenic mice which also have increased Cyclin D1 expression (Lee et al., 2000). Researchers have also suggested involvement of Cyclin D1 overexpression in parotid gland carcinogenesis (Pignataro et al., 1998). The mechanisms involved in the synergistic e€ects between p53 and c-neu in salivary glands is unclear. Oncogene

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1452

Figure 6 Expression of p53 and p21 genes, aneuploidy and Cyclin D1 analyses. (a) Northern blot of p53 and p21 gene expression in normal glands and tumors of transgenic mice. Normal salivary gland (lane 1), salivary gland tumor (lane 2), breast tumors (lanes 3 and 6) from c-neu; p53+/7 mouse (lane 1). Salivary gland tumor from a c-neu;p537/7 mouse. Tertocarcinoma and sarcoma from two p537/7 mice (lanes 4 and 5) and normal breast from wild type female (lane 7). (b) Graph demonstrating aneupolidy found in parotid gland tumors from c-neu;p53+/7 mice. Immunohistochemical detection of Cyclin D1 in breast (d) and parotid gland (e,f) tumors from c-neu;p53+/7 mice. (c) Negative control. Arrows point to Cyclin D1 nuclear staining

However, these studies suggest that p53 loss in conjunction with c-neu overexpression quickly converts the normal status of the parotid gland into a transformed state indicating that p53 inactivation is a tissue speci®c early event in parotid tumors. However, in breast tissue, p53 alterations did not increase the rate of breast tumor formation. Therefore, p53 can exhibit early or late e€ects on tumorigenesis that is presumably dependent on tissue speci®c modi®ers. However, additional events including increased proliferation associated with Cyclin D1 overexpression and aneuploidy are also required for tumorigenesis. Our data suggests this phenomenon may be caused by the Oncogene

pleiotrophic functions of p53 and is complicated by interactions of many complementary or inhibitory tissue speci®c factors. Materials and methods Generation of p53-deficient mice that carry the c-neu oncogene p53-de®cient mice generated by gene targeting (Donehower et al., 1992) were in 129 background and the transgenic strain, TG.NK, that carries an activated c-neu oncogene driven by an MMTV enhancer/promoter (Muller et al., 1988) was in FVB background. To minimize background variation, the

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

p53-de®cient mice were outbred three times into FVB background before being mated with the TG.NK strain. To genotype o€spring from these crosses, we extracted genomic DNA from 1 cm sections of tails using a salt extraction method. The DNA was digested with a fourfold excess of EcoRI or BamHI, and electrophoresed on 0.7% agarose. Following Southern blotting, the nitrocellulose ®lters bearing EcoRI digested DNA was hybridized with a 32P-labeled p53 cDNA probe to detect the mutated allele of the p53 gene. The nitrocellulose ®lters bearing BamHI digested DNA were hybridized with a 32P-labeled speci®c to the SV40 poly(A)+ to detect the presence of the c-neu transgene (Muller et al., 1988). Based on the combination of the c-neu oncogene, wild type and the mutated alleles of the p53 gene, the animals were divided into ®ve groups; p537/7 c-neu;p537/7;p53+/7; cneu;p53+/7 and c-neu;p53+/+. Animals de®cient in p21 were genotyped as described (Deng et al., 1995). Histology, antibody staining and TUNEL assays Animals in all experiment groups were checked weekly after birth and when the ®rst day tumors were noticed the animals were recorded as developing tumors. Complete autopsies were performed and both gross and microscopic evaluations were performed. Tissues were ®xed in the omni®x, blocked in paran, sectioned, stained with hematoxylin and eosin, and examined by light microscopy. Cyclin D1 antibody was purchased from Santa Cruz Biotechnology. Detection of primary antibodies was performed using the ZYMED Histomouse2 SP Kit according to manufacturer's instructions. TUNEL assays for apoptotic cells on tissue sections were carried out as recommended by the manufacturer (Oncor). Chromosome preparations from tumor cells Chromsomes of tumors were prepared following the approach described by EP Evans (Evans, 1987) with minor modi®cations. Mammary and/or salivary tumors were dissected, brie¯y washed in PBS and minced in MEBM/ MEGM media (Clonetics) containing 10 mg/ml of bacterial collagenase type III (mammary tumors) or type II (salivary tumors) (Gibco). After overnight digestion, released cells and

®nely minced tissues were pelleted, washed in PBS and cultured in 5% DMEM overnight at 378C (5% CO2). The adherent cells were then treated for 1 h with Colcemid (Sigma). The media was then removed and the cells were trypsinized and resuspended gently in 0.56% KCl and incubated at 378C for 5 min. The cells were then ®xed in a 3 : 1 mixture of methanol:acetic acid and the cell were then processed as described.

1453

Northern blot RNA was isolated from tissues using RNA TeT-60 based on the protocol suggested by the manufacturer (Pharmicia). About 20 mg of total RNA from each sample was electrophoresed on the 1% agarose gel and transferred to a Gene-Screen ®lter. The ®lter was then hybridized with a 32Plabeled probe speci®c to the SV40 poly(A)+ in the c-neu transgene to detect the expression of the transgene. For the detection of the expression of the p53 and p21, 2 mg of poly (A)+ RNA were loaded on each lane. The probe for the p21 is a 430 base fragment containing the exon 2 of the gene (Deng et al., 1995). The probe for the p53 is the full length cDNA (Muller et al., 1988). In situ hybridization In situ hybridizations were carried out using standard procedures. Brie¯y, tissues to be analysed were ®xed in freshly prepared 4% paraformaldehyde in PBS for 4 h at 48C and embedded in paran and sectioned at 4 mm. Slides with sections were brought to room temperature, dewaxed and hybridized to a 35S-UTP-labeled RNA probe speci®c to the cneu transgene overnight in a humidi®ed box at 528C. Hybridization bu€er, washing and RNAase treatment. Slides were dipped in emulsion (Kodak NTB-2) and exposed for 4 ± 15 days before developing.

Acknowledgment We would like to thank D Shanmugarajah for technical assistance.

References Bargmann CI, Hung MC and Weinberg RA. (1986). Cell, 45, 649 ± 657. Bouchard L, Lamarre L, Tremblay PJ and Jolicoeur P. (1989). Cell, 57, 931 ± 936. Brugarolas J, Chandrasekaran C, Gordon JI, Beach D, Jacks T and Hannon GJ. (1995). Nature, 377, 552 ± 557. Califano J and Eisele DW. (1999). Otolaryngol. Clin. North Am., 32, 861 ± 873. Cardi€ RD, Sinn E, Muller W and Leder P. (1991). Am. J. Pathol., 139, 495 ± 501. Dasika GK, Lin SC, Zhao S, Sung P, Tomkinson A and Lee EY. (1999). Oncogene, 18, 7883 ± 7899. Deng C, Zhang P, Harper JW, Elledge SJ and Leder P. (1995). Cell, 82, 675 ± 684. Donehower LA. (1997). Cancer Surv., 29, 329 ± 352. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery Jr CA, Butel JS and Bradley A. (1992). Nature, 356, 215 ± 221. Earp HS, Dawson TL, Li X and Yu H. (1995). Breast Cancer Res. Treat., 35, 115 ± 132.

el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell, 75, 817 ± 825. Elson A, Deng C, Campos-Torres J, Donehower LA and Leder P. (1995) Oncogene, 11, 181 ± 190. Evans EP. (1987). In Monk, M. (ed.) Mammalian Development. IRL Press Limited, Oxford, pp. 93 ± 114. Graus-Porta D, Beerli RR, Daly JM and Hynes NE. (1997). EMBO J., 16, 1647 ± 1655. Hall PA, Hughes CM, Staddon SL, Richman PI, Gullick WJ and Lemoine NR. (1990). J. Pathol., 161, 195 ± 200. Hand AR, Pathmanathan D and Field RB. (1999). Arch. Oral. Biol., 44 (Suppl 1), S3 ± 10. Harvey M, McArthur MJ, Montgomery Jr CA, Bradley A and Donehower LA. (1993). FASEB J., 7, 938 ± 943. Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Thiagalingam S, Kinzler KW and Vogelstein B. (1997). Mol. Cell., 1, 3 ± 11. Hunter T. (1991). Cell, 64, 249 ± 270. Oncogene

p53 and erbB2 synergy in salivary gland tumorigenesis SG Brodie et al

1454

Oncogene

Hynes NE. (1996). J. Mammary Gland Biol. Neoplasia, 1, 199 ± 206. Isola J, Chu L, DeVries S, Matsumura K, Chew K, Ljung BM and Waldman FM. (1999). Clin. Cancer Res., 5, 4140 ± 4145. Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi S, Bronson RT and Weinberg RA. (1994). Curr. Biol., 4, 1 ± 7. Jimenez RE, Wallis T, Tabasczka P and Visscher DW. (2000). Mod. Pathol., 13, 37 ± 45. Kamio N. (1996). Virchows Arch., 428, 75 ± 83. Knudson CM and Korsmeyer SJ. (1997). Nat. Genet., 16, 358 ± 363. Knudson CM, Tung KS, Tourtellotte WG, Brown GA and Korsmeyer SJ. (1995). Science, 270, 96 ± 99. Lee JM and Bernstein A. (1993). Proc. Natl. Acad. Sci. USA, 90, 5742 ± 5746. Lee RJ, Albanese C, Fu M, D'Amico M, Lin B, Watanabe G, Haines III GK, Siegel PM, Hung MC, Yarden Y, Horowitz JM, Muller WJ and Pestell RG. (2000). Mol. Cell. Biol., 20, 672 ± 683. Levine AJ. (1997). Cell, 88, 323 ± 331. Linke SP, Clarkin KC and Wahl GM. (1997). Cancer Res., 57, 1171 ± 1179. Lucchini F, Sacco MG, Hu N, Villa A, Brown J, Cesano L, Mangiarini L, Rindi G, Kindl S, Sessa F, Vezzoni P and Clerici L. (1992). Cancer Lett., 64, 203 ± 209. Malkin D, Jolly KW, Barbier N, Look AT, Friend SH, Gebhardt MC, Andersen TI, Borresen AL, Li FP, Garber J and Strong LC. (1992). N. Engl. J. Med., 326, 1309 ± 1315. Menard S, Tagliabue E, Campiglio M and Pupa SM. (2000). J. Cell. Physiol., 182, 150 ± 162. Morgan SE and Kastan MB. (1997). Adv. Cancer Res., 71, 1 ± 25. Muller S, Vigneswaran N, Gansler T, Gramlich T, DeRose PB and Cohen C. (1994). Mod. Pathol., 7, 628 ± 632. Muller WJ, Sinn E, Pattengale PK, Wallace R and Leder P. (1988). Cell, 54, 105 ± 115. Oswald F, Lovec H, Moroy T and Lipp M. (1994). Oncogene, 9, 2029 ± 2036.

Pignataro L, Capaccio P, Carboni N, Pruneri G, Ottaviani A, Bu€a R and Broich G. (1998). Anticancer Res., 18, 1287 ± 1290. Pinkas-Kramarski R, Alroy I and Yarden Y. (1997). J. Mammary Gland Biol. Neoplasia, 2, 97 ± 107. Pinkas-Kramarski R, Shelly M, Glathe S, Ratzkin BJ and Yarden Y. (1996). J. Biol. Chem., 271, 19029 ± 19032. Polyak K, Xia Y, Zweier JL, Kinzler KW and Vogelstein B. (1997). Nature, 389, 300 ± 305. Press MF, Pike MC, Hung G, Zhou JY, Ma Y, George J, Dietz-Band J, James W, Slamon DJ, Batsakis JG and ElNagger AK. (1994). Cancer Res., 54, 5675 ± 5682. Resnitzky D and Reed SI. (1995). Mol. Cell. Biol., 15, 3463 ± 3469. Rice DH. (1999). Otolaryngol Clin. North Am., 32, 875 ± 886. Samuelson LC. (1996). Annu. Rev. Physiol., 58, 209 ± 229. Schulze A, Zerfass K, Spitkovsky D, Henglein B and JansenDurr P. (1994). Oncogene, 9, 3475 ± 3482. Sherr CJ. (1996). Science, 274, 1672 ± 1677. Shih C, Padhy LC, Murray M and Weinberg RA. (1981). Nature, 290, 261 ± 264. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A and McGuire WL. (1987). Science, 235, 177 ± 182. Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A and Press MP. (1989). Science, 244, 707 ± 712. Stal O, Sullivan S, Sun XF, Wingren S and Nordenskjold B. (1994). Cytometry, 16, 160 ± 168. Tonin PN. (2000). Semin. Surg. Oncol., 18, 281 ± 286. Tzahar E, Waterman H, Chen X, Levkowitz G, Karunagaran D, Lavi S, Ratzkin BJ and Yarden Y. (1996). Mol. Cell. Biol., 16, 5276 ± 5287. Varley JM, Evans DG and Birch JM. (1997). Br. J. Cancer, 76, 1 ± 14. Vogelstein B and Kinzler KW. (1992). Cell, 70, 523 ± 526. Webster MA and Muller WJ. (1994). Semin. Cancer Biol., 5, 69 ± 76. Yin Y, Solomon G, Deng C and Barrett JC. (1999). Mol. Carcinog., 24, 15 ± 24.